I am trying to create a machine learning model in a country which has high inflation. With this model, I am trying to predict the price of a second hand car. As my train data, I have second hand car sales from 2014 and 2015. As my test data, I use car sales from 2017 and 2018.

As features, I am using the attributes of a car: Horse Power, Engine Size, KM, Vehicle Type, Brand, Model... In addition, I am using 2 econometric features: Special Consumption Tax and Euro Currency.

These two features are positively correlated with the price of the car. If Euro Currency increases, second hand prices increase. It is also true for special consumption tax. You can see the change in Euro Currency for the last 5 years.

Euro Currency

I am using XGBoost to create model. However, RMSE and MAE are very high. Should I change the algorithm? Or should I use more econometric features?

  • $\begingroup$ I'd suggest you take a loog at LSTM models. This link shows an implementation. $\endgroup$ – Lucas Farias Apr 2 at 20:38
  • $\begingroup$ Have you tried a standard VAR approach to begin with? $\endgroup$ – usεr11852 May 2 at 21:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.